This proposal requests funding to extend the follow-up of the Bypass Angioplasty Revascularization Investigation (BARI) patients to a minimum of 10 years. BARI is a randomized international multicenter clinical trial that compares a strategy of initial percutaneous transluminal coronary angioplasty (PTCA) to that of initial coronary bypass graft (CABG) surgery for selected patients with multivessel coronary artery disease (CAD). The original hypothesis of BARI was that a strategy of initial PTCA is no worse than one of initial CABG when assessed by mortality at 5 years. The trial began in July of 1987. Investigators from 18 sites randomized 1829 patients. In addition, 2013 patients who were eligible but not randomized were recruited into a registry. Overall results of the trial indicated no significant difference in 5- year mortality between an initial strategy of CABG vs an initial strategy of PTCA. However, in the subgroup of patients receiving medical treatment for diabetes, 5-year survival was significantly better for those assigned to CABG. An ancillary study of cost and quality of life at 7 centers revealed that over 5 years, PTCA was cheaper than CABG by an average of 2620 dollars. Data entry and management will be performed centrally at the Coordinating Center and abridged versions of the current data collection forms will be used. Site coordinators will continue to be involved with patient contact and thus the high rate of follow-up achieved at five years (98 percent) is expected to continue. Patients are just now entering the period during which their original surgical grafts are beginning to fail. Since the majority of CABG patients received mammary artery grafts (82 percent), BARI will be able to validate recent findings of improved long-term patency with these grafts. However, it is unknown how failure of additional saphenous vein grafts will affect outcome. Thus, an adequate comparison of the CABG and PTCA strategies must include the 5 - 10 year follow-up period. Extended follow-up of the diabetic subgroup will be particularly important because the treatment difference is expected to widen. The registry will continue to serve as an important mechanism to assess the generalizability of the randomized trial, while also providing a source of additional patients for specific subgroups. With the release of 5-year data, the most productive period of the study in terms of publications has just begun. The BARI infrastructure must be preserved so that the BARI database can be used to its fullest potential.